Agent 3.0.5 unstable on VMWare ESX 3.0.1

[remi_hyperic]

I installed the Hyperic agent 3.0.5 on 12 VMWare ESX 3.0.1 servers. 3 agents died after a couple days with no errors in the log files. If I restart them they run for a few minutes or hours and then exit again the same way. After the agent died I found a file "hs_err_pidXXXX.log" (see attached file).

All 12 ESX servers are installed the same way.

[dougm_hyperic]

I haven't seen this before, but from what I can gather the agent is running low on memory, a garbage collection is triggered.. something within the VM object has gone stale by the time finalize() gets called. I will look into this, a couple of possible bandaids in the meantime:
Add the following to agent.properties, so only the vmware plugin is loaded:

plugins.include=vmware

Use the jre that is also serving the esx web console, rather than the bundled HQ jre:

HQ_JAVA_HOME=/usr/lib/vmware/webAccess/java/jre1.5.0_07 ./hq-agent.sh start